KTH oorg kemi

juli 1959

Some laboratory methods

in current use at the department of inorganic chemistry,

Royal Institute of Technology, Stockholm 70, Sweden.

page

Preface 1
=======

General Laboratory notes, Water, Glassware. Removal of co2 and o2• 2

=======

Recrystallization,

Weights, Volumetric glassware, 3

�� � ���� ��� ��1����
= =

Acids and bases

===============

Preparation and purity. HCl04, HCl, HN03 5

Na2co3, NaOH, KOH, Ba(OH)2

Analysis, Acid-base-titrations. Strong acids. Strong bases, 8

Carbonate. KHC039 Tl2co 3•

Ionic media
= = =========

Preparation and purity, Na0104, NaCl, NaBr, LiBr, KN03, Ba(Clo ) •

4 2
9
Analysis, NaC104, Ba(Cl04)2, other ionic media. 11

Other stock solutions

=====================

Preparation. AgClo4. Other metal perchlorates. 11

Analysis, Metal, Perchlorate, Silver- halogenide titrations, 12
Redox titrations,

Emf work
========

Thermometers and potentiometers, Thermometers, Potentiometers. 13

Wilhelm and other glassware,

Electrodes for [H+]. General, H 2• Quinhydrone. Glass. 15

Redox and metal electrodes, Redox. Mercury and amalgam, Silver, 18
Reference electrodes, Calomel. Ag+/Ag. �
X-/AgX,Ag.X-,HgX -/Hg, 20

Determination of H and h. Correcting H, Equations for E. Gran dia- 21

gram, Calculating E , j, and h. H in so­

o
lutions of metal salts. Determination of

E • E for metal, amalgam, and redox elec-

o 0

trodes,
 Preface
=======

The present notes are based on the experience of workers in this

laboratory up to June 1959o They will be revised from time to time

in the future; suggestions for additions or improvements will be most

welcome.

For the type of work we have chosen a high standard of accuracy

must �e maintained. This little collection of methods is'not a sub­

stitute for the personal instruction of new co-workers, nor for ex­

perimental judgement. It goes without s�ying that nobody need feel

obliged to follow the procedures described here if he wants to try

other methods. This manual may be useful in bringing attention to

certain good standard methods9 and also in encouraging discussions

and -the testing of new methods.

It is important to be accurate and careful in the right place.

Therefore you should try to realize what is the accuracy of your various

determinations� and how possible errors may influence the final results.

Double checks must be made, if possible using methods as different as

possible, so that errors in reading or mixing ( which may happen to

anyone ) cannot get as far as the final results. Still worse are syste­

matic errors, such as impurities in the reagents used, or systematic

errors in the analytical methods. You may remember that the determina­

tion of total perchlorate with a cation exchanger, which we once consi­

dered reliable, seems to introdMce systematic errors in the presence of

some polyvalent cations.

The methods given above are certainly not perfect,and may be im-
oy
proved. If you have information�xperience which differs from the

statements made here, or which you think ought to be brought to the

attention of others anyway, will you please let me know so that it can

be introduced into the next editiono Thank you.

Lars Gunnar
 - 2 -

General
========

Laboratory notes� please keep your notes in such a state that

someone else ( such as yourself three years later ) can understand what

the notes mean and find what he wants. It is especially important that

the primary data ( analyses, calibration�, emf-data ) which form the basis

for all the following work, are kept in such a way that they do not dis­

appear or are not mixed with secondary calculations. Preferably use a

bound book, or numbered loose leaves, with your name,,and date of the

experiment on them. Never rely on your memory; its half-life is shorter

than you think.

Water. Distilled water should be tested for impurities that may

influence.measurements� co , silicate, Cl • co is removed by boiling9

2 2
or by bubbling through N j free from co o Chloride is detected with
+
2 2
Ag ( compare with knovvn concentrations ) . Silicate is detected by the
yellow molybdate test ( yellow color on additionafNH and NH molybdate
3 4
Nils, Georg II)o More sensitive (too sensitive for some purposes) is

the molybdate-benzidine test ( Vogel, Qualitative Analysis 1945, p. 292,

first method, Nils ) . For some purposes ( e.g. rinsing bottles and under

fortunate circumstances, even mixing equilibrium solutions ) , distilled

water from the tap may be used; in general it is safer to use the doubly

distilled water from Frans Egs room. Even that must be tested!

Glass- ware. For acid solutions use Jena Gerate or Pyrex which,

however, will dissolve in the course of time to give a few tenths of a

mM Si ( OH ) • Alkaline solutions should be kept in polythene bottles

4
with thick walls ( which let through less co and H o vapor ) . Usually
2 2
the screw stopper is tight; this can be tested by squeezing the bottle

a litfle, while screwing the stopper into position; if the bulge is

still there next time you use the bottle, you can be certain the stopper

is tight. Rubber stoppers will stain the polythene and probably conta­

minate the alkali; therefore, screw stoppers or smooth polythene stoppers

are preferable. ( Nils )

For short periods5however, one may keep equilibrium solutions
-
with not more than 0 , 1 M OH in Jena Gerate glass vessels ( Gosta,

anions III, A C S 1959, 722). In the course of a titration, only some

0.01 mM Si will dissolve, and glass has many advantages, for instance

it is possible to see a faint opalescense.

 - 3 -

·Work with alkaline solutions is discussed under "sodiwn hydroxide"

below.

Carbon dioxide is removed from air or N by passing the gas

2
through a U-tube containing ascarite; some ascarite is easily carried

with the gas as a fine dust, and may be detected by sniffing the gas.

Therefore a plug of cloth or cotton must be placed in the U-tube down­

stream, and it is desirable to wash the gas with H 0 free from co2•
2
To exclude oxygen, nitrogen from cylinders _is used; the oxygen

it contains (which may be about 0.1%) is usually removed by passing

°
through a tube containing activated copper on kieselguhr, at 180 C

(F R Meyer and G Ronge, Z angew Chern �(1959)637). Do not forget to

reduce the copper oxide from time to time by passing H gas over it.
2
The active copper is usually prepared by freshmen following Brauer,

Handbuch der praparativen Chemie, p. 353. - The same catalyst removes

oxygen from H by catalytic combustion.

2
To remove o from H , commercial "desoxo" (Pt black) may be used
2 2
also. It has the advantage that it can be used at ordinary temperaturss,

and so ca n be left overnight without danger. The speed of the gas stream

must not be too high (Nils).

The gas (N or H ) may be purified further by passing through

2 2
wash-bottles containing 10% H2so4 and 10% NaOH (Nils et al, anions II,

A C S 19579 1041) , and finally bubbled through a wash-bottle containing

the ionic medium to be used.

In any recrystallization one should give the impurities a good

chance to get away with the motLer liquor. The difference in solubility

at two temperatures is used9 and only seldom is an evaporation necessary.

The mother liquor should be removed as completely as possible; as little

as possible should remain occluded or on the surface of the crystals .

A last rinsing may be desirable (compare NaC10 , Kurt's method below).

4

Weights are calibrated by Richard's method (Kalthoff-Sandell 1952,

p. 227) every year (Georg II), or whenever one finds that one's weights

have been exchanged (Ake, Nils).

A standard set of weights will be kept by Nils.

 - 4 -

Volumetric glass-ware (measuring flasks, pipets, burets) are

calibrated by weighing with boiled water at 25°C. Good procedures are

given in Kalthoff-Sandell 1952, p. 227, and in Kalthoff-Stenger II,

1947, p. 3, correction factors may be taken from Landolt-Bornstein,
Hw I� Pe 78 or Kalthoff-Stenger II, p, 22. For a 50 ml or 20 ml buret,

every 5 ml is weighed, for a 10 ml buret, every 3 ml. Pipets are weigh­

ed three times (Sirkka).
The errors are often several tenths of a %; for small pipets the
errors are larger. It is worthwhile to select pipets and burets that
have a negligible correction (the others may be used by the freshmen).
When alkaline solutions are used, it is important to check the volumes
often,
The water weighings have been compared with 5 M NaCl (Sirkka),
3M NaC10 (Ake)� 3M NaCl (Hele�), 1 M KN0 (Helen), Hg (Helen).
4 3
The same results were obtained if the burets were emptied slowly (at
most 1 em/minute, Helen; 10-15 minutes between the readings, Ake).
If pipets are emptied freely, 3M NaCl gives about 0.1 %less volume

than H 0 or 1 M KN03 (Helen).

2
Pipets should be allowed to drain for 1 min. after emptying,
(which takes �t min.). The volume is then reproducible to better than
0.05 % (Georg II). In the calibration they should be emptied in such
way that one is not obliged later to use a very time-consuming method
of emptying; for instance, let the solution run out from the pipet,
blow out the last drop, and take the pipet away (Ake). Georg II� This
is not recommended for other workersg blowing tends to make the volume
less reproducible, and may contaminate the solutiono

With all readings of burets, pipets, and flasks, a piece of

white translucent paper with a piece of black opaque paper on the lower
half is used to give a sharp, black reproducible meniscus (Helen).
Cleaning of glass-ware� dichromate +sulfuric acid (Georg II,
Nils), c Hc1 trichloroethylene (which dissolves the stop-cock grease
2 3
apiezon N) and afterwards strong KOH in ethanol (Helen). Afterwards
distilled water and ethanol; suck the buret dry with filtered air (lke).
As usual: use own judgement.
 -
5
-

Acids and bases

===============

Pre�aration and purity

�.E£�ric acid, commercial pro analysi quality is about 70 per

cent and at present does not seem to contain iron(III) or chloride, so

that it can bo diluted and used directly for our purposes, It should

always be tested for chloride, if this is important; compare the opale­

scence with Ag+ and solutions of known log [Cl- ], so that a maxi mum

value for log [Cl-] can be given. If impurities are found in the c ommer­

cial product, it is better to try a new bottle, or a new brand. It does

not seem advisable to re-distil it in the lab ,

Hydrochloric acid is purified by distilling the az eotrop i o (con­

stant-boiling) solution. It is sometimes important that Br- is absent

e . g . even s;:naJ.l amounts of Br- may be concentrated in AgCl, or Hg c1 ,

2 2
as a solid solution, and give erroneous solubilities and emf, One may
-
test for Br using eosin (E Glintelberg, diss Copenhagen 1937, p, 94)1

the method works but requires some time and patience (Nils).

The B�- tast of E Schulek (Anal Chim Acta �(1948)74) is preferable

(Georg II). It depends on the reactions. 2 Br- + 2 C l + 2 CN- �

2
- - +
2 BrCN + 4 Cl -) BrC�J + 2 I + H � I2 + Br- + HCN. Example. Bring
1 ml HCl + 1 d�op of methyl red into a stoppere d 20 ml ,Erlenmeyer flask.

Neutralize with NaOH. Add 3 - 4 drops of 01 water (saturated, freshly

2
prepared) !i v,rai t 2 - 3 min. and then add 3 - 5 drops of 5 % KCN and a big

crystal (� 4 m�of KHC0 ,

) Acidify with a few drops of (same) HCl,
3
(co expels the air), and add 3 - 5 drops of 2M KI. Titrate the r
2 2
' formed with l rr.'1 Na s2o • Do not forget the blank test.
2 3
Bromide is r-::nnoved from HCl according to Gtintelberg (Diss 1937,

p. 95): to 20% HCl about ten small portions of KMno are added, and
4
the Cl is boiled of£ completely after each addi ti on ; each time taking
2
with it the major part of the remaining Br • Finally the HCl is dis­
2
tilled.
Merck 38 % HCl pro analysi always seems to contain small amounts

of iron(III).

N�ig_§Lcid, pro analysi, can be dilu te d and used without fur ther

purification (Helen); since it often contains considerable amounts of

Si(OH) , �istillation of a c o nsta nt-boil i n g acid is preferable (Georg II).

4
 -6-

Sodium carbonate, Merck pro analysi, ·contains varying amounts of

C l , often about o. 01 %.
- Select a bottie with little Cl- (Ake). It
°
should be re-crystallized; at.50- 60 0 it is slighty less soluble than
° °
at 30 C but dissolves more quickly. Cool to 0 0, It is important that

small crystals are formed, since in large crystals,mother liquor is often

occluded. By careful work one may reduce the concentration of Cl to

about 1 /100 of the original value in a single re-crystallization {Georg II).

Sodium hydroxide. In a polythene bottle, 50% NaOH is prepared

from solid NaOH pro analysi and water (Sorensen 1909 see Kalthoff­
Stenger II, p. 70). After a few days, the sparingly soluble Na co has
2 3
sunk to the bottom of the bottle. {In case of emergency, the 50% NaOH,

which will be filtered anyway, may be used 1 - 2 hours after mixi ng).
The 50% NaOH is filtered through a G 4 glass filter into a auction
flask. The filter must never be sucked dry! When a sufficient amount
of filtered 50% NaOH has been obtained in the suction bulb, the suction

and filter are removed, and a pipet is introduced, A white skin of

Na co is always formed on the 50% NaOH, so the tip of the pipet should
2 3
be thrust through the skin. (Do you ever use your mouth when sucking a

pipet?)
In the meantime, a polythene bottle has been almost filled with
boiled distilled water, and a stream of N (free from c o ) is bubbled
2 2
through it or passed over the surface. When the pipet has been filled
J (this is not the time for measuring volumes!), the lower part of th�
pipet, and its tip are wiped carefully. (If the 50% NaOH inside the
pipet is opalescent, reject it). Now transfer the 50% NaOH to the
bottle containing water, with the N stream on, but leave some solution
2
in the pipet. Stopper the bottle and leave it in a thermostat room for
one day before analyzing. (Nils).
The NaOH solutions should never be poured from this bottle, since
this may introduce carbonate (Nils). To transfer the solution to a
buret or pipet, take off the stopper, fix on the old polythene tubing
with N2 stream as before (see sketch!), dip the tip of the buret (pi:pet)
into th e solution and suck, through an asoarite tube if you like.
To transfer the alkali to a measuring flask, use a bure t, never pour
directly! (Nils).
 - 7 -

su ck
4l!:

�,�-1;.(
buret
·� ·�

ascarite

N2 from cylinder

rubber
tubing
ascar it e

Usually substances are dissolved from polythene bottles which make

the solutions "oily" and even give them a faint smell. These substances

do not seem to influence the titrations but of course are not desirable.

Less trouble is obtained with polythene bottles that are old or have been

soaked with warm alkali.

Potassium hydroxide. Commercial solid KOH pro analysi is dissolved

in water9 and a small excess of Ba(OH) is added to precipitate the car­

2
2+
bonate ionso The solution is decanted or filtered, and excess Ba re­
+
placed by K in an ion exchanger (Dowex-50) which has been saturated with

K+ (KCl) and carefully washed. The inside of the ion exchange tube should

be coated with paraffin or polytheneo In the KOH solution, one must have

[OH-]� 1M, otherwise the ion exchanger is seriously attacked. The

eluate is collected in a polythene bottle under it should be tested

N�;
2+ -
from time to time for Ba with H2so49 ·and for co3 with BaC1 • (Nils)
2

Barium hydroxide. Commercial Ba(OH) pro analysi contains consi�

2
derable amounts of Cl and must be recrystallized twice. It is then

dissolved in co -free water9 filtered and diluted. If the solutions are

2
 - 8 -

then treated with the s2me precautions as NaOH solutions, no opalescence

of Baco is obtained, giving some confidence in the method (Ake)�

3

Analysis

In acid-base titrations with a color indicator such as methyl red

wherever possible5 the base should be in the titration flask, and the

acid in the buret. Acid is added till the color is changed, then the

solution �s boiled and cooled with tap-water (on outside of flask!).

If the basic color h.:.:�s returned, more acid is added, until the color is

again acidico The solution is boiled, cooled, if necessary more acid

added etc, until the acidic color remains on boiling. For [OH-] <:.10 mM9

an error of 0.5 % is easily made if this procedure is not usedR Even

stock solutions may be analyzed, with good accuracy, by emf titration

with, for instance, a glass or quinhydrone electrode and Gran-diagrams.

Strong acids. Titrate against weighed KHC03 (all), check with

T12co (Georg II)9 HgO + KI (Erik I). Checks must agree within 0.1 %.
3
Georg II prefers T12co , which has a high equivalent weight and can also
3
be used for standardizing redox solutions (iodate, cerium(IV), and per­

manc;ana te) 0

Str�ng baseso Titrate against HC104 (Georg II) or weighed hydra­

zine sulfate (N2H Hso4). Nils also adds excess HCl to a known volume
5
of NaOH solution, evaporates, and weighs the NaCl formed. Weighing the

NaCl is a good check, but should never be used as a primary analysisg

3+
t it requires that the NaOH used be free from Cl-, Fe , and other impuri­

ties.

According to Georg II hydrazine sulfate is all right, but takes

time to prepare' KH(Io ) can be prepared pure only in an atmosphere

) 3 2
complotely free from dust (which is hard to achieve here) and KH-phtha­

late gives a large carbonate error if one does not titrate in hot solu­

tion.

A final check is obtained from the acid-base titrations in the

equilibrium experiments (see "emf work" below)o

The amount of carbonate in the alkali can be estimated by titrating

with HCl, first with phenolphthal�jn (pH = 8, HCO� ), and then with methyl

red (pH = 5, H co ). Alternatively use an emf-titration with a Gran

2 3
+
diagram (distance between the H and OH- lines at zero height). A more

elaborate apparatus for expelling the C02 and absorbing it in Ba(OH) is

2
described by Gosta (ACS 1959, 724).
 - 9 -

Georg IIg we should determine acids, bases, and some redox solutions

by coulometry. An accuracy between 0.1 and 0.01 % may be obtained, and

time may be saved. A silver coulometer + a color indicator would suffice.

Pqt�um hydrogen carbonate. Merck pro analysi seems to contain

Q001 % Na as sole impurity; after one re-crystallization ( from water, not

)
ot�12��1�1 the product does not even give a positive flame test for Na

( G9org II )o It should be kept for several days in a desiccator in C02

Gtmosphere ( Winkler 1915), which converts small amounts of K co

2 3
to

�rco3• Sirkka has silica gel in the desiccator, Georg II has H so •

2 4

T h�::rr ( I )carbonate contains some lead II ( ) � two crystallizations

in aPt bowl gives a purity better than 99.99 %.

Ionic media
===========

Preparation and purity

4
Socium perchJorate is usually prepared from perchloric acid and
-
solium carbonate ( free from Cl , preferably re-crystallized ).
Concentrated HCl0 is added in small portions to a mixture of solid
4
Hn co and water; all the Na co need not be dissolved - the solubility
2 3 2 3
°
is only �-5. 5 c; /
1 00 ml at 1 00 C. A\ small excess of HClO
4
is added; as

usual in preparative work, one tests by transferring a drop on a glass

rod to an indicator paper and not by adding indicator to the solution.

T�o C0 is boiled away, the solution is neutralized with NaOH using

2
phnnolphthale:_n ( to a faint pink ) or bromocresol-green +methyl red ( to a

g�ey color ) o Prepare your indicators yourself; those mixed for the fresh­

men often contain alkali which will spoil your analysis. ( Georg II, Sirkka )

Alternative� Concentrated HClo is stirred mechanically in a beaker,

4
and Na co added spoonwise until a small excess has been added. Then a
2 3
small excess of HC10 is added, the solution is boiled and finally neutra­
4
li �ed with NaOH to pH ·=-3 8 ( phenolphthalein ) . ( Nils, Ake ) .
The solution is allowed to stand for a week if possible. If Fe,

Al, or Si are present, they precipitate. Filter through a G 4 filtar,

acidify with HC10 , boil, adjust pH again with bromocresol green-methyl

4
rod. The solution is boiled in a beaker and covered with a watchglass

( otherwise, a crust of solid NaCl0

4
is formed along the walls, which may

easily be decomposed to NaCl by local overheating . )

 - 10 -

Sometimes a precipitate is obtained at this stage, which must then

be removed by filtering. The solution is now boiled until, at 137 C, a

0

skin of NaC10 forms on the surface of the solution. Cool the solution
4
to 60°C while stirring it vigorously; NaCl0 H o precipitates and is filt­
4 2
ered off by suction. After one more re-crystallization (from 137°C to

60°C), the product is usually satisfactory. If the solution is cooled

only to 80°C, anhydrous NaC10 is obtained. (Nils, Ake)

4
(Sodium perchlorate was earlier prepared by re-crystallization of

the commercial salt from ethanol-water mixtures. However, some ethanol

is easily oxidized in the drying process, and even small amounts of im­

purities (like acetate) in the ionic medium can cause serious errors.

Kurt has recently made satisfactory NaC10 by double crystallization of

4
the product of G.F, Smith (which seems to vary in quality between various

batches) in water. When as much mother liquor as possible has been sucked

from the crystals, (still in the funnel), they are rinsed with a small

amount of chloride-free distilled water. In this way [Cl-] in the stock

solution, which would otherwise have been around 1 mM, may be reduced to

0.02 mM (Kurt).)

Impurities. If correctly prepared, 6 M NaC10

contains less than
4
0.01 mM Fe3+, Cl-� or SO�- (Georg II), The concentration of Si is at

most 0.1 or 0.2 mM (Georg II),estimated from the residue after double

evaporation with HCl; Nil? found no residue if the salt was doubly re­

crystallizedQ Georg II does not recommend the silica-molybdate method�

) it gives a positive test even with distilled water, and the extinction

drifts with time.

The concentration of AlJ+ increases with time. After a month,

distinct Al lines are observed in the spectrograph; the concentration

is about 0.01 mM; The Al comes from the glass, which is of low quality

for big bottles. It would be desirable to test our solutions more often

spectrographically; this can be done with Axel J�s apparatus (Georg II).

In this way, unpleasant surprises may be avoided.

Sodium chloride, Merck pro analysi, is heated to 120°C (Sirkka)

200°C (Nils) or 360°C (Georg II), placed in a desiccator, weighed, and

used, It may contain a basic impurity (carbonate?, a few 0.1 mM in 3M

NaCl), which is neutralized with HCl, As usualg make your own indicator,

and beware of the freshman alkaline indicators. Sirkka has found that

NaCl is more basic after drying at }60°C, but +his is denied by Georg II�
 - 11 -

The NaCl may be re crystallized from HCl. To the saturated solu­

-
tion at 0°C, HCl gas is added (from cone HCl +H so4 dropwise). This,
2
however, does not seem necessary.

It is also possible, but does not seem necessary, to prepare NaCl

from Na co and HCl, in the same way as NaClo4•

2 3
Br- impurity has been less than 0,01 %in recent years, If it is

desired to remove the last traces of Br, the solution may be saturated

at 0°C with Cl (from KMn04 and Br-free HCl), which is then boiled off
2
with most of the Br • This is repeated a few times (Glintelberg, diss
2
1937, p, 95). Test for Br-, see HCl.

Sodium bromide, Mallinckrodt pro analysi, (no Si, or carbonate)

is re-crystallized once (Nils)�

Lithium bromide, Kebo, re-crystallized twice.

·�
\ ..

Potassium nitrate, pro analysi, re-crystallized twice.

Barium perchlorate is prepared from barium hydroxide and HC104•

Analysis

Sodium perchlorate solution. A known volume of solution is evapo-

o
rated in an oven at 120- 125 C to constant weight. Alternatively the

solution may be weighed, and the density determined using a pyknometer·

(more accurate with strong, viscous solutions, Sirkka). A known amount

of solution may also be passed through a H+-saturated cation exchanger,

and the H+ClO � in the el�ate determined by titration.

Barium perchlorate: evaporate in an oven at about 200°C and weigh

the anhydrous Ba(Clo4) • The desiccator must contain a strong water ab­
2
sorbant such as Mg(Clo4) •
2

Other ionic mediag LiCl, NaBr, KNo , are determined by similar

3
methods: evaporating and weighing, or ion exchange.

Other stock solutions

=====================

Preparation

For our purpose it is not always necessary to prepare saits in a

crystalline form. It is enough to have a solution of known composition.

Silver perchlorate .• is given as an example (it is often used in

external or internal reference electrodes)

 - 12 -

a) Add NaOH solution to a hot AgN0 _solution, and wash and decant
3
+
the precipitated Ag 0 about 20 times. Small amounts of Na remain, but
2
for many purposes this is not serious. An excess of Ag o is added to a
2
known amount of HClo , the solution is boiled till it seems saturated,
4
and then separated. The pH of the solution is then 4 - 5 (lke, Ido).

b) Evaporate AgN0 + HC10 two or three times in aPt dish.

3 4
Unfortunately, some AgCl is formed by reduction with dust particles

(Georg II would like to have a dust-free room somewhere).

The preparation of other metal perchlorates is described in our

various papers; for a new problem, it may be useful to read the liter2.-­

ture carefully, but very often one must invent a method for oneself, +c

obtain a satisfactory purity. Perchlorates may be prepared from HCl0 +

11_
2+ 3+ + 2+ �
metal oxide (Pb , Bi , UO �(add H o ), c u ) , metal + Cu(Clo ) (sn2- ),
2 2 4 2
+ 2+ 3+
evaporation of metal nitrate with HCl0 (Ag , Pb , Fe ), evaporation
4
3+ 3+ iV
of metal chloride + HC10 (AI , Ga , Zr ): electrolytic oxidation or
4
3+ 2+
reduction from another oxidation state (Tl , Fe ).

Analysis

The metal concentration should be determined by two independer:.t

methods. The perchlorate concentration is determined approxioatoly b�y­

+
passing the stock solution through a H -saturated ion exchanger a�d
+
titrating for H C10 . 4 With polyvalent metal ions this method often

gives systematic errors (the ion exchangers seem to retain some 010 4
which is very hard to wash out); the accuracy suffices to adjust the

ionic medium, but not for determining H. Sometimes the accuracy may bo

increased by first precipitating the metal ion, for instance with }i'1!3,

separating and washing the precipitate, passing filtrate + wash water

+
through a H -a.aturated ion exchanger, and titrating.

Earlier, attempts were made to remove the metal hydroxide by pre­

cipitating with NH , evaporating the filtrate + washings, and W3iGhinr,

3
NH c1o ; however, this salt is somewhat volatile, so that low results
4 4
are obtained (Sirkka). Precipitating the metal hydroxide with, s::.y,

NaOH in hot solution and back-titrating the excess NaOH does not give

sharp end points. Attempts have been made to determine 010 4 directly
3+
by reduction with Ti in concentrated H so , or by precipitating vdt�
2 4
+ -
(C H ) P Cl and weighing, but errors of -0.5 to -1.0% were obtained
6 5 4
(Georg II).
 � 13 -

It seems preferable to determine ClO � approximately using the

ion-exchange method, but to get the final value for H by emf-titration

with a Gran diagram, in the course of the final equilibrium experiment.

Silver-halogenide titrations. Use silver electrode and Gran diagram

+
( plot both Ag and X- lines ) . Primary standard� NaCl is best ( Georg II,
Nils ) , Ag may get a skin of Ag S ( Otto! ) , and AgN0 may contain some Ag
2 3
metal, Ake finds Mohr titration convenient, Nils and Georg II prefer

emf + Gran.

Redox titrations. Pt or Au electrode, Gran diagram ( or just the

steepest jump ) . As primary standards Georg II prefers T1 co or KI03

2 3
( preferably with ICl titrationvisual end-point Brennecke et al, Neuere

t' massanalytische Methoden, Stuttgart 1951, p. 122)

.SJ

Emf work
========

Thermometers and potentiometers

Thermometers. We have two thermometers calibrated by Technische

° °
Reichsanstalt in Berlin� one for 15 - 25 C with 0.01 divisions, one
° °
for 0 - 50 C with 0.1 divisions ( Georg II). The thermostat temperatures

should be checked more often than is perhaps generally done9 and tempera­
°
tures should be kept correct to better than 0.05 C.

If a standard thermometer is placed in the thermostat, at the place

where the equilibrium solution will be, it is possible to obtain a con­

°
stancy of better than 25.00 ± 0,01 C ( Helen ) . - It should be remembered

} that the heat of reaction caused by each addition in the course of a

titration may change the temperature, and subsequent adjustment of tP�-­

perature may be seen as a drift in the emf.

On the suggestion of Georg II, a resistance thermometer has been

ordered from the National Bureau of Standards; when it arrives, it should

be checked against ice, Na so etc, and then the other thermometers

2 4
should be calibrated.

Potentiometers. The compensators have often been compared.

The Leeds and Northrup instruments differ by less than 0.01 mV for total

E up to 100 mV. The Cambridge potentiometer has been compared with one

1 J. N; at"high E, differences of up to 0,01 mV have been observed.

 - 14 -

The valve potentiometers should be compared with the compensator,

(which is assumed to be correct) at least once or twice a year. There

are slow changes in the emf of the little built-in cadmium cell, and of

the zero point of the scale. (Georg II). The deviations from the com­

pensator are often up to 0.3 mV for high emf, and for high and low read­

ings on the 100 mV wheel (Ake).

The standard cells available seem to keep the E on their labels to

within� 0.02 mV.

Wilhelm and other glass-ware, The three mimeographed sheets with

drawings were made for our freshmen at KTH1 the texts are in Swedish,

which may amuse some of our guests.

Bild 7 is the standard "Wilhelm" apparatus (Wilhelm, Sirkka and

' \
VII
Lars Gunnar, ACS 1952, 901) with reference electrode, salt bridge, and

titration vessel, (in this case set for a freshman titration). Usually

the same ionic medium is used throughout the cell, and the salt bridge

contains the pure medium, without any additions,

After some thought you will find out how to use the stopcocks and

glass joints in order to fill the apparatus with the solutions required.

It is normal to spoil one titration by opening the wrong stopcock.

If you do it more than twice or break more than one glass electrode,

watch your hands more carefully. We can afford the solutions and elec­

trodes, but think of your time!

Bild 6 is a shorter Wilhelm or perhaps rather a Walter (Nernst)

for cruder work. Note the double inlet for H2.

Bild 1 - 4 are simplified drawings of electrodes in the forms used

by our freshmen •

.t
 A 1.

_i5 n:i.:

Pt-bleck

. . ·.�
Fast kinhydron
·---..--
. ..... ..� ----···�

Glas

Bild 3

Pt-tr�d

Ag-belaggn

AgCl-kri st.
 � r £: (')

_I
( !

A ,--·---- (-�-'
� I

-�-0 -

��}
e
)
/"•· ·· ---
B
II

iI
::::.
/
_ .. --
H2 1· n
- = '
-··. ,
.\,-1
-,
':.c:- .�- -;-'- �2 '_) � ��
.. §
' /-('--.
.· ' ..
_.;_ .

//)
.

11
n tl

c:CJ ft II
'
'
- ;
I
Avlo�psrOr fOr gasen

J�j =t J: �.: ' ! ",li� -� t b �- 17·.·;:-··

- ;-::- =---
''·
(I -·
,
a _ .. Byrett
ij L
�J

Il l
t

'·.I
I

Vatskelank cnasL:Jl. ekt.-l:orl

·-- ·��

: i£MJ L-.
Vatgaselek­
trod

R'HC �-, _ t -t �::�

Ql'���=w_
__

'T'i ll eclningsr<.;.c
for gu.nb:ub111; ng

�
Bifd 6.
 A 1. bild 7.

c Forradslosning for NaCl(a\)

I Byrett
H TilloppsrOr fOr N2 �
G Elektrod Ag!Agi -�"'
FOrradskarl for NaH03(aq ) ���'-,
1
F
,_I =·
E Avloppsror,

i
)
j
l
Jl
1:

L,'A ---���-=lektrod Ag,A�l

B Elektrodkarl for NaCl (aq
�....:�.

--------�----------�--��

D Trevagskran for va:�skelan�-�n NaCl(aq_)/NaNO;}(a:�;,)

J Kapillar for vatskelanken NaN03(aq)/provlosning

K Frovlosning
L Flerhalskolv (titrerkarl
 - 15 -

At present we measure h With hydrogen, quinhydrone, and glass. elec­

trodes; one day, perhaps, we ·will make Ag 0-Ag-electr�des for use in

2
alkaline medium. In theory the hydrogenelectrode is best but often it

cannot be used since it reduces some substances in the equilibrium solu­

tion; in addition, it comes to equilibrium more slowly than the other

two.

The quinhydrone electrode can be used only in acid solution

(log h > -7), since quinhydrone is destroyed in alkaline solution.

In strongly acid solution correction must be made for the basicity of

quinone (Georg II, A CS 1956, 1340). Unfortunately it is possible that

hydroquinol or quinone may form metal ion complexes; Clara has found evi­
r'i ,; 2+
. dence for this with Cu , (Arkiv Kemi 2(1956)175).
The glass electrode

gives lower accuracy since a valve potentiometer must be used; the theory

is not too well understood, and E may be influenced by other ions that

are adsorbed on the glass. It is desirable to compare at least two dif­

ferent types of electrode.

Hydrogen electrode. A rectangular Pt foil is welded to a Pt wire

(a H torch is preferable for the welding; city gas may give contamina­
2
tion with s, and electrode trouble), The Pt wire is sealed through a

closed glass tube, and connection on the inside is made with a copper

wire, or Hg +copper,

In earlier experiments with badly buffered redox solutions, E was

found to drift if the s olution could leak in through small cracks in the

glass around the Pt wire, For this reason, the Pt wire was made long,

and the glass tube was kept well above the solution (LG, SKT 1946, 56).
" Nowadays this does not seem necessary; if trouble arises, it might be
I

worthwhile to try this method.

Pt coil may also be used. If a H torch is used, E reaches its

2
final value more rapidly, and different electrodes agree still better.

(Nils),

The Pt foil (wire) is cleaned in cone HN0 (or aqua regia for a
3
s hort time), water, and ethanol, and then ignited in an ethanol flame,

Then it is coated with Pt, Pd, or Ir black by electrolysis,

a) Platinum·black: dissolve 1 g PtC14 in 30 ml H o, add 0,1 ml

2
0,1M HCl, and 20 mg PbAc •
2
Electrolyse (as cathode) 1 - 3 minutes with

0.3- 0.15 A (Sirkka), 0.4- 0.2 A (Bates, pH 1954, p.1 66 - 167),

 - 16 -

2
0,01 A/cm (Georg II). Sometimes, no black cover is obtained; it then
helps to add more PbAc (Sirkka). Finally the foil is electrolysed (as
2
cathode) in 0,1 M H so (Georg II), 10% H so4 (Sirkka, Nils) for 20 - 30
2 4 2
mi nutes with the same current density as before, When the last electro­
lysis was cut down to 5 minutes, equally good results were obtained
(Sirkka). The lead acetate does not influence the electrode (Georg II).
b) Palladium black: 1 g PdC1 + 14 ml cone HCl + lOO ml H o.
2 2
Times and current densities as for Pt, same electrolysis in H so4 as for
2
Pt; (Hamer and Acree, 1944; Bates pH 1954 p, 168, Sirkka), It seems
� that the composition of the solution can be varied considerably, Nils:
2
1 g PdC12, 50 ml cone HCl, 50 ml H2o, 2 min 0,2 A/cm , 20 min in 0,1 M
2
H2so4, 0.5 A/cm • If the electrode is used a second day, the electro-
.� lysis in H2so4 should be repeated on the morning of the second day,
, 2
c) Iridium black: 1 g IrC1
4 + 20 ml 1 M HCl 0,001 A/cm in 12 h
(Georg II). More rapid equilibrium, less influence of traces of o
2
(Georg II); just as good (or bad) as Pt or Pd in alkaline solutions
(Nils), (Lewis, :Brighton, Sebastian, JACS 1917, 2249),
!Jhe electrode is rinsed with distilled water before it goes into
the equ:.:.librium solution. Some also wipe it off with filter paper,
others do not,
Old Pt black is removed with aqua regia; when this has been re­
peated a number of times the Pt surface becomes etched and dull, and it
is time to reject it, (Nils),
The hydrogen gas must be made free from oxygen, using activated
Cu or "desoxo",
For each new titration, a new electrode should be made, -The
H stream should be divided into two (regulated by a pincher): one bubb­
2
les through the electrode tube, one enters below the surface of the
solution, Before the titration, first bubble o -free N for at least
2 2
30 mi n, then H (o -free) for about an hour or until changes in the rate
2 2
of bubbling (always slow!) do not influence E, The outlet for H should
2
be under water, and the titration vessel should be tight, and not un­
necessarily large, It may be a good idea to have the apparatus standing
with N or H bubbling overnight, after a titration, and to check E the
2 2
next morni ng (Nils).
,

In alkaline solutions which do not c;ontain o or other oxidizing

2
substances, equilibrium is obtained within 30- 60 minutes for each

point (Nils),
 - 17 -

In acid solutions, free from 0 , the first equilibrium takes about

2
60 min, and the following 20 - 30 min each (Georg II), E is not influ­

enced by the rate of bubbling H if it is between 3 and 60 bubbles per

2
minute (Georg II).(20- 30 is a normal rate). In the presence of tri­

positive ions, the equilibrium is reached slowly, perhaps because of

adsorption on the metallic surface (Georg II).-

Quinhydrone electrode, The quinhydrone must be free from iron(III)

and excess acid, and have the right Q/QH ratio, Crystallize from water
2
and dry over H so in a vacuum at room temperature (Georg II); crystallize
2 4
°
� from boiled water but never let the temperature exceed 70 C (Helen),

Purity tests: acid is detected with indicator in the saturated

-
aqueous solution; iron(III) is detected with SCN after destroying the

quinhydrone by boiling with concentrated H so ,

:• 2 4
The electrode is a bright Pt foil (see H electrode), which is
2
rinsed by boiling with cone HN0 , and ignited in an ethanol flame, The
3
ignition is repeated before each titration, and also, as a check, from

time to time during the titration,

With the quinhydrone electrode, the solution must be stirred me­

chanically and not with N ; in the latter case E drifts because quinone
2
evaporates (Georg II, Helen),

Glass electrode, Each glass electrode has a personality of its

own, and black sheep are often encountered, No mark seems to be comple­

tely reliable, No glass electrode can be used over the whole pH-scale;

if you try to determine pK by an acid-base titration, even the best ones

w
will give errors of a few tenths of a unit, If you let a glass electrode

make a long return trip along the pH-scale, you will find that its E
0
has changed, Here are a few observations:

"Blomgren electrodes" were .excellent in acid solution (pH 3 to 0,4)

but are hard to obtain nowadays (Ake),

Radiometer yellow (for.acid solution) is good for pH> 3 (Ake),

between 2 and 7 (Helen), more rapid than the red electrode in acid solu­

tion (Helen).

Radiometer red ("alkali resistant") may be used also between pH 7

and 1,5 (Sirkka, Helen, Ake) but will not stand strongly acid solution

for a long time. Neither the yellow nor the red one is good between
pH 5 and 7 (Ake), -�he red ones are often very good in the range pH =

6 to 12 (�ils), However, the E should always be checked before and

0
 - 18 -

after a titration, which is rejected if the difference is too large.(Nils).

Jena electrodes always seem to creep in alkaline solution (Nils);

as good as Beckman (Georg II).

Beckman acid, good up to pH 9 (Ake), 6 (Georg II); lives for a year

at 25°0, a month at 50°C; even when it has started to drift at 50°0, it ��

may give good results at 25°0 (Georg II).

Beckman alkalineg looks fine, but too little experience yet (Nils).

Storageg acid electrodes in 0,1 M HOl; if the electrode is stored

in distilled water, it takes several hours to reach equilibrium (Georg II).

A new electrode is kept in 0.1 M HOl for a few hours, after that an equal

time in distilled water; repeat this several times, and then let it stand

over night in distilled water before use, Store in distilled water (Helen),

Glass electrodes seem to be destroyed by standing in UO + solution; �

';,'• �
after a night in a UO + solution, they are not fit for precise work.

The same may be true for other cations!

E should be determined before and after each titration (both, if

0

possible) without removing the electrode from the equilibrium vessel.

It seems that E may drift slowly in one direction, rather than jump,
0

sometimes up to 7 mV in a single week (Sirkka), If the electrodes are

taken out, washed, and then placed in the same solution, E usually keeps

to within! 0,2 mV at 25°0; at 50°C, greater care is needed because of

electrostatic charging� metal parts must be earthed etc (Georg II).

Redox and metal electrodes

Redox electrodes. Pt foil as described under "hydrogen electrode",

q,v, for possible influence of cracks in the glass tube,

For each new system it may be worthwhile to try a number of diffe­

rent "inert" metals, and if the equilibria are slow, one may even add

some redox catalyst (I I- , indigo sulfonate etc, but in small amounts

2 +
only and preferably not at all).

Bright Au and Ir-Pt (25% Ir) foils may also be used, They are

treated like Pt; clean by boiling for a short time in aqua regia (Georg II),

or by boiling in 1 g1 HN0 and igniting in an ethanol flame (Sirkka).

3
2
Well-behaved redox systems (Hg +, Hg +; Tl+, � 3 2
Tl +; Fe +,
3
Fe +;

QH , Q, H+) usually work equally well with bright Pt and bright Au. For
2
2 4 .
the more tricky system uo2 +, U + , H +, br�ght Pt, Au, or Ir-Pt were slow
 - 19 -

2
(equilibrium only after more than hours), but platinized and especially

palladiatedPt (preparation, see H2 electrode) gave the equilibrium E

4+ 3+
rapidly (Sirkka). With ce , Ce , bright Au-electrodes were somewhat

more rapid than brightPt; on platinizedPt, cerium(IV) was reduced in a

few minutes (Sirkka). - It may be worthwhile in some cases to try Hg­

electrodes; however, the solution must not be able to.oxidize Hg metal

to Hg
+
�or to precipitate a metal that forms an amalgam. Each new system

gives a wide field for imagination and judgement.

Store bright metal electrodes in 0,1 M HCl if they are used daily'

for long interruptions, they are better stored dry,

Mercury and amalgam electrode, Often the Hg is used as a pool on

the bottom of the vessel9 and contact is made with aPt-wire sealed

through a glass tube. This electrode behaves very well (Sirkka). Note
•
that a skin of water can form on the Pt-tip, which gives trouble with

calomel electrodes,

Other ways to get metallic contact� Pt electrodes sealed from below

through the vessel, Hg pool in cup (Pt should never touch the water!).

C�pillary with Hg dipped into the pool; Pt electrode sealed through the

capillary just below the tap, C apillary of J-shape, ending in a little

cup,

Amalgam electrodes can be made in similar ways,

The amalgam can usually be prepared by electrolysis of a suitable

salt solution (Ake Pb,Bi; Georg II Tl, In) and also by direct reaction
2
of metal and Hg. Pb is brought into contact with Hg under 0,0 5 M HC104

(Ake); it may then be stored under the aqueous solution of HC10 and
4
finally forced through a tube to the equilibrium vessel, by applying N2

pressure on the surface,

For dry preparation of an amalgamg remove the oxide layer with a

•

knife, place the metal and Hg in separate arms of an inverted U-shaped

vessel, wash with co2 and evacuate twice, Shut the taps and pour the Hg

over the metal (In, Sn, Pb, Georg II). The amalgam is then transferred

to the equilibrium vessel after the vessel and all connections have been

filled with co2, in flowing co2 atmosphere,

Amalgam that has been prepared electrolytically may be sucked (or

forced) over to the U-shaped vessel where the amalgam is stored after the

"dry11 preparation, It is then transferred as described above.

 - 20 -

Silver (chloride) electrode.s t3.re prepared according to A S Brown,

JACS 2£(1934)646. Pt coil of length about 1 em, one end sealed through

a glass tube, is coated with Ag by electrolysis from 1 % KAg(CN)2 for

5 - 6 h with at most 0.3 rnA. The solution should not contain free HCN.
Wash carefully with water, and then electrolyse for about 30 min in 0,1 M

NaCl, with the same current, this time as anode; AgCl then replaces the

outer (e,nd perhaps irregular) layers of Ag.

Nils prefers rnA for 8 h, and then electrolysis in dilute HCl,

Br f ree eccording to Guntelberg.

- If 3 - 4 electrodes are made, they may

be compared with each other •

•
The .AgCl layer must not be too thick, it is enough if 5 - 10%

of the Ag is converted to AgCl (Georg II). These electrodes are used

+ +
not o nly in Cl-, but also in Ag solutions, since they respond to Ag
• more rapidly and reproducibly than Ag foil, or Ag-coated Pt •

In some solutions, AgCl is not desirableg in strong alkali it is

converted to Ag 0; it dissolves in strong Cl solutions. It may be con­

2
venient sometimes to use an Agi coating, which is obtained either by

electrolysis in Nai solution instead of in NaCl, or by dipping the chlo­

ridized electrode in about 1 M Nai solution overnight, Agi also dissolves

in strong iodide solutions.

If it is necessary to measure emf to! 0,02 mV or better, oxygen

should be excluded from the solutions.

Reference electrodes

A number of different reference electrodes have been tested here

in �he course of time, and tastes have varied with time and worker,

Nowadays calomel electrodes seem to be used only for very crude

measursmonts, They are hard to make and they give emfs that are not too

reproducible� a drift of 1 - 2 mV has been observed, The error may be­

come much larger, if the Pt electrode loses its contact with the Hg

( water skin on Pt), Of course the electrode is also sensitive to small

amounts of Br-.

Silver ion-silver electrodes have been kept constant to 0,1 to

0,2 mV (Sirkfa) or 0,02 mV (Georg II). They are convenient in acid or

neutral perchlorate or nitrate medium, If a glass electrode is used and

+ +
the syste� does not react with Ag or Ag, one may have Ag in the equili­

brium solution and thus do without a salt bridge and liquid junction
 - 21 -

potential (Georg II).

For alkalin e soJ.'.ltoi ons .. OH-/Ag 0,Ag and OR-/HgO,Hg have been used
2
in other pl ac e s but do not seem to have been tried here, On the other

hand, h alo g enide electrodes have been tested, which can be used both in

acid and alkaline soluti o n :

Sil ver-halo.g_Gno�entate electrode, usually Cl-/AgCl,Ag or rather

Cl-,AgCl /Ag.; Silver chloride is rather soluble in strong chloride me­

dium, perhaps 10 m]! Ag in 3M Cl-. To get a reproducible potential one

should ei the r se.turate the reference electrode solution with AgCl, or at

least dissolve; a knovm ar.:ount of AgCl in it. If the solution is not

saturated, tho smc.ll amount of AgCl on the electrode surface will dis­

solve, but it has alre ady fulfilled its purpose by smoothing the elec­

trode surfaceo The s at urated electrodes have been found to vary by 0.5

• to 1.0 mV; hcwevor, no serious attempt seems to have been made to keep

them constant si�ce one has hitherto been most concerned with changes

in E and not with absolute values (Sirkka).

With Br- or I :r.::::dig one may make the corresponding Ag electrodes:

AgBr and Ae;I are n.lso soluble in excuss of halogenide,

Mercl..:!,;gr-halog_�no�purate electrode, At high concentrations of

halogenide, the IIc;(I) :1a:ogenides are d.ecomposed: Hg Br + 2Br- -tHg(l)

2 2
+ !
HgBr - etc� On0 uses a clear solution with a halogenomercurate(II),

and me tall i c merot :--y. .

1!. preliminary study of such electrodes indicates
- 2- .
that B r , H�Br 4 ;H� lS promising, and that also the corresponding iodide
0 '·:.;)

elect rode m�y be uc0do (Sirkka, Nils).

•

Gener�l on emf measurements# for lady chemistsg if you want to

avoid a sh o c k from you� apparatus, don't wear nylon stockings and a nylon

underskirt in dry weather (Helen),

Determination of H and h

Correct iJ!g_�� The previous analysis will give a rather good appro­

ximation for H in each solution used in an equilibrium experiment.

For a final Rnalyni8, c�e may use an emf acid base titration in connec­
-
tion with tho equilibTium studies, Since the error is small, it is not

important vrhich .
s -.J..':lti0n one assumes to be correct; the calculations

often come out sc:-rlc::wha � ;;impler if the error is assumed to be in the

original solution (s) =� the vessel, whereas the solution (T) in the buret

is correct.
 - 22 ....

Wg uations for E,With an h-sensitive electrode in the equilibrium

solution, we have (for not too large h )( Georg II and LG, Arkiv Kemi 2_(1953)425)

E = E + 59.15 log h - jh ( acid side ) (1 )

0
E = E0a - 59. 15 log (OR-] + j
a
(OR J (alkaline side ) ( 2)
2
where and j
E, j are constants. (
If the electrode in the equilibrium
o a
solution is the minus pole, the signs are reversed . )
Usually, j has the same value for all electrodes in each medium

( with glass electrodes, small individual deviations may be encountered)o

E may vary somewhat with time and therefore should be determined, if

0
possible, both before and after each titration. If the experiment is

started with a titration without any metal ion present ( thus, only ionic

medium, acid and base ), one obtains E ( for the day in question , ) j, and
0
a check on the concentration of the acid and base solutions,

Gran diagram� ( Analyst ll(1952)661) are applied to the titration

of a strong acid in the following way. The volume v of solution T that

has been added is plotted against a quantity which is proportional to the

amount
+ remaining in the solution.
of R
This quantity is

v antilog )/59.15) const (v0+v)[R+] (3)

(v
0+ ) ( (E-E' =

0
where (v +v ) is the total volume of the solution after the addition,
0
E is the measured emf, E' is an arbitrarily chosen constant; if the
0

h-sensitive half-cell is the negative pole, the exponent should have the

opposite sign. This plot s hould give a straight line which intersects
+
the v-axis at the equivalence point, v ' where all H would have been
t
neutralized,

It may be worthwhile, in very accurate work, to find the equivalence

point from the Gran function in some more elaborate manner� calculating

the intersection point numerically, or assuming approximate values for

v and the slope and studying a deviation function.

t
For alkaline solutions, a corresponding function is proportional
-
to the remaining amount of OH namely

In applying Gran diagrams on the acid side one may instead plot the con-

centration H
appr-cx,
( calculated from the preliminary analyses ) against

antilog
o
((E- E')/59,15), which gives the error in H. For a simple titra­
 - 23 -

tion the result is the same as above, and this form is necessary if

samples of the solution are withdrawn with a pipet. (Ake).

Calculating E , j, and h. When v (or the error in H) has been

0 t
determined from the Gran plot 9 one may calculate the corrected values

for H = h in each point of the acid-base titration, If E- 59.15 log h

is plotted against h, the result is usually a straight line, which gives

E and j. Small deviations at low h can often be remov�d if the correc-

o
tion for the concentrations is changed slightly (still within the accuracy

of the Gran diagram). The same is true for calculating E , and j on

oa a
the alkaline side,

If E and j are known, the E measured for any solution (also those
0
containing metal ions) gives h, using equation (1 ). At low values of h,

a two-step approximation suffices and sometimes one may even neglect jh,

At higher acidities it is practical to �ake an auxiliary plot of ('E0- jh)

as a function of E to get a good first approximation.

H in solutions of metal salts may be determined in a similar way

by acid-base titration and Gran diagram, preferably using solutions with

the right ionic medium, and a few different total metal concentrations, B,
When hydrolysis of the metal becomes noticeable, a bend is observed, but
2+
well-behaved metals like Pb , zn
2+ +
�
, and also UO , give little difficulty,

A rather precise method has been given by Clara (Arkiv Kemi 2(1956)175)
2+
and applied to Cu -solutions, The idea is to titrate first with NaOH

at rather low acidities (but still no hydrolysis), then to add known

amounts of an acid of a rather high concentration, The E is obtained

0
from the latter part of the titration, the small correction to H from

the first part. For details, see Clara�s paper, Ake has proposed a

similar procedure in which it is not necessary to calculate E between

0
2+
titrations (lke, paper on Pb ).

Determination of E , when all concentrations are accurately known,

0
can be made as described above by a preliminary acid-base titration with

no hydrolyzing species present, If one starts with a weakly hydrolyzing

2+
metal ion in the solution (like Pb ), one may start the titration at

high values for B( > 10 mM) with H _,-:-J20 mM, and add NaOH of known concen­

tration; this gives E and a check on H. At low values for B one should
0
not start with large values for H, since the hydrolysis will then be only

a small part of H, and Z will be less accurately known, It is then better

 - 24 -

to start with a low value for H, and increase z. After this main part

of the titration, a known amount of acid is added (to about 0.02 M), and

E is calculatedo
0
For stronglyhydrolyzing metal ions it is necessary to introduce

some kind of stepwise approximation. See for example the second paper
• 3+ 4+
on Bi (Ake, ACS 1959, ); or Th in self mediumg (Sirkka and Lars

Gunnar, ACS 1959,533).

To find E0 for metal� amalgam, and redox electrodes, one may use
• the same value for j obtained earlier for h-electrodes. (Note the devia­

tion in the quinhydrone electrode at high acidities). (Georg II, ACS

1956,1340). Here also it is desirable to work in a range where the hydro­

lysis can be neglected; otherwise E must be obtained by stepwise approxi-

o
mation.

For work in alkaline solution similar formulas can be used, only

now the liquid junction potential is proportional to [OR-], See for in­

stance Nils et al, ACS 1957,1034,

Before you use Gran's method, some modification of it, or any

other mathematical or graphical method, you should understand how the

formulas were derived, what are the underlying assumptions, and whether
I
these are applicable to your case or not.